Maistrenko, Valentin I.

Relevance. Today there are no real models of the human brain's heuristic activity. In this regard, now it is absent real models of heuristic activity of the human brain, according to the accumulated experimental material, the possibilities of modeling two basic principles of the operation of real neural networks of the human brain was presented.

Goal. The proving of real neuroemulator possibility for system synthesis as a model of heuristic behavior of human brain.

Methods. We propose the numerical repetition of electroencephalogram registration (EEG) and construct the special matrix of their (EEG) comparison. For neuroemulator we propose the numerical repetition and choice value of weight of coefficient w_i0for any diagnostic parameters x_i(like for real EEG).

Conclusions. It is shown that such two procedures (reverberation and chaos) are identical to the heuristic activity of the brain. The brain selects the main x_i* from the set of xi variables and at the same time we move from the phase space m to the new phase space with little dimension n (n<<m). The choice of the main x_i*is a problem of system synthesis (finding of the order parameters). It is realized at the expense of reverberation and chaos to neural networks. We must mark: the artificial neuron nets emphasize demonstrate such behavior at every time.